To enhance peak-rates within a cellular technology, multi-carrier or carrier aggregation solutions are known to be efficient. Each carrier in multi-carrier or carrier aggregation system is generally termed as a component carrier (CC) or sometimes is also referred to a cell. The term carrier aggregation (CA) is also called (e.g. interchangeably called) “multi-carrier system”, “multi-cell operation”, “multi-carrier operation”, “multi-carrier”, “multi-frequency carrier” transmission and/or reception. This means that the CA is used for transmission of signaling and data in the uplink and/or downlink directions. One of the CCs is the primary carrier or anchor carrier and the remaining ones are called secondary or supplementary carriers. Generally the primary or anchor CC carries the essential UE specific signaling. The primary CC exists in both uplink and direction CA. The network may assign different primary carriers to different UEs operating in the same sector or cell. Further, carriers may be activated or deactivated for different UEs. Thanks to carrier aggregation, the UE has more than one serving cell: one primary serving cell and one or more secondary serving cell. The serving cell is interchangeably called as primary cell (PCell) or primary serving cell (PSC) or serving cell on primary CC. Similarly the secondary serving cell is interchangeably called as secondary cell (SCell) or secondary serving cell (SSC) or serving cell on secondary CC. Regardless of the terminology, the PCell and SCell(s) enable the UE to receive and transmit data. More specifically the PCell and SCell exist in DL and UL for the reception and transmission of data by the UE. The remaining non-serving cells on the PCC and SCC are called neighbor cells. The CCs belonging to the CA may belong to the same frequency band (aka intra-band CA) or to different frequency band (inter-band CA) or any combination thereof (e.g. 2 CCs in band A and 1 CC in band B). The carriers in intra-band CA can be adjacent (aka contiguous) or non-adjacent (aka non-contiguous).
Dual-Carrier High-Speed Downlink Packet Access (DC-HSDPA, also known as Dual-Cell HSDPA) was introduced within the 3rd Generation Partnership Project (3GPP) Rel-8. DC-HSDPA enables reception of data from two cells simultaneously, transmitted on two adjacent carriers in a same base station and sector, to individual wireless devices. The concept of DC-HSDPA is in 3GPP Rel-10, being extended e.g. to 4 downlink carriers (known as 4C-HSDPA).
In HSPA release 10 up to 4 downlink carriers can be aggregated as 4C-HSDPA where the downlink carriers or cells may belong to the same frequency band or may be split over two different frequency bands e.g. 3 adjacent downlink carriers in band I (2.1 GHz) and 1 downlink carrier in band VIII (900 MHz). In HSPA Rel-11 even up to 8 downlink carriers may be aggregated, this configuration may be denoted as 8C-HSDPA; the downlink carriers may be distributed over 2 or even more bands.
To complement DC-HSDPA, in 3GPP Rel-9, Dual-Carrier High-Speed Uplink Packet Access (DC-HSUPA) was also introduced. DC-HSUPA enables an individual wireless device to transmit data on two adjacent carriers simultaneously to a radio access network. DC-HSUPA according to 3GPP Rel-9 is in essence an aggregation of legacy (Rel-8, single-carrier) HSUPA.
In LTE (Long Term Evolution) intra-band CA up to 5 downlink carriers and 5 uplink carriers each of up to 20 MHz may be aggregated by the wireless device. In LTE inter-band CA, up to 5 downlink and 5 uplink carriers each of up to 20 MHz and belonging to different bands can be aggregated by the wireless device. Even additional carriers may be introduced in future releases. CC in CA may or may not be co-located in the same site or base station. For instance the CCs may originate (i.e. transmitted/received) at different locations (e.g. from non-collocated BSs (Base Stations), RRHs (Radio Remote Head) or RRUs (Radio Remote Unit).
Although the additional spectrum bandwidth associated with multi-carrier operation does not increase “spectral efficiency” (maximum achievable throughput per cell per Hz [bps/cell/H]), the experienced user data rates are increased significantly. In particular, for bursty packet data traffic at low and moderate load, the data rate is proportional to the number of carriers exploited. Moreover, power inefficient higher order modulation schemes can be avoided (which is especially important in the uplink) and the practical as well as theoretical peak data rate of the system are naturally increased.
In a network according to the 3GPP specifications a RNC (Radio Network Controller) controls radio resources and radio connectivity within a set of cells. Handover and radio access bearer admission control is presumed to be conducted in the RNC based on measurements of path loss etc on a primary carrier (alternatively referred to as an anchor carrier). RAN (Radio Access Network) is according to the 3GPP specification responsible for the radio transmission and control of the radio connection. A Node-B, also referred to as Node B, handles the radio transmission and reception within one or more cells. In case of a distributed RAN architecture where Node-B and RNC functionality as defined in 3GPP specifications are co-located in the base station, the base station would naturally handle also these functionalities. In a DC-HSUPA capable Node-B, the other carrier, which is referred to as a secondary carrier, is assumed to be configured by the RNC for a given DC-HSUPA capable wireless device and then scheduled and activated by the Node-B whenever feasible and useful (with the standard objective function to maximize the supported traffic volumes, or aggregate system throughput, subject to fairness criteria and quality of service constraints, such as minimum bit rate or maximum latency requirements). A primary carrier, on the other hand, may not be temporarily deactivated by the Node-B: to deactivate a certain primary carrier for a connection, the connection is either released, or an inter-frequency handover is performed (in which case another carrier will become the primary carrier).
For each wireless device connected in DC-HSUPA mode, the serving Node-B hence controls whether or not a secondary carrier is activated, and a separate grant is selected for each activated carrier.
Furthermore, if a secondary carrier is activated by the Node-B, it is assumed that the Dedicated Physical Control Channel (DPCCH), which includes a sequence of pilot bits, is transmitted in uplink on that carrier, and the Node-B hence tries to detect this signal.
UTRAN (Universal Terrestrial Radio Access Network) is a collective term for the Node B's and RNCs which make up a UMTSRadio Access Network (RAN). The wireless device may be in a CELL_FACH state, where the UTRAN may redirect the wireless device to another frequency. In a future system, one can envisage multi-carrier operations in the CELL_FACH state. Node-B controlled carrier selection of the uplink transmissions will then be introduced.
Load estimation in the WCDMA (Wideband Code Division Multiple Access) uplink is performed for many reasons in prior art. Most importantly, the present scheduling of enhanced uplink traffic is based on the principle to schedule users until a load threshold is reached. Such scheduling decisions are taken every 2/10 ms transmission time interval (TTI). Thresholds are typically used in order to maintain a planned coverage, and to maintain cell stability avoiding inner loop power control (ILPC) rushes. When coverage is addressed neighbour cell interference is incorporated in the load measure, this is not the case when cell stability is treated. The scheduling and load estimation functionality and algorithms are both located in the WCDMA RBS.
It is also possible to use the estimated uplink load in load based admission control algorithms. Also this is known in prior art. These algorithms use the uplink load in order to determine if new users can be admitted in specific cells. The admission control functionality is located in the RNC node. Signaling means for signaling of load is available over an NBAP interface. It is e.g. shown in H. Holma and A. Toskala, WCDMA for UMTS—Radio Access for Third Generation Mobile Communications. Chichester, UK: Wiley, 2000 that without advanced interference suppressing (IS) receivers and interference cancellation (IC), the load defined at an antenna connector is given by the noise rise, or rise over thermal, RoT(t), defined by
            RoT      ⁡              (        t        )              =                  RTWP        ⁡                  (          t          )                    N        ,where N is the thermal noise level as measured at the antenna connector. The definition of RTWP(t) is the total wideband power
            RTWP      ⁡              (        t        )              =                            ∑                      k            =            1                    K                ⁢                                  ⁢                              P            k                    ⁡                      (            t            )                              +              I        ⁡                  (          t          )                    +      N        ,also measured at the antenna connector. Here Pu(t), u=1, . . . , U, denotes the power of uplink user u, and I(t) denotes the power as received from neighbor cells of the WCDMA system. A problem that now needs to be addressed is that the signal reference point is, by definition, at the antenna connector. The measurements are, however, obtained after an analogue signal conditioning chain, in a digital receiver. The analogue signal conditioning chain may unfortunately introduce a scale factor error of about 1-3 dB. Fortunately, all powers of the cell are almost equally affected by the scale factor error so when the RoT is calculated, the scale factor error is cancelled as
            RoT              Digital        ⁢                                  ⁢        Receiver              ⁡          (      t      )        =                              RTWP                      Digital            ⁢                                                  ⁢            Receiver                          ⁡                  (          t          )                                      N                      Digital            ⁢                                                  ⁢            Reciever                          ⁡                  (          t          )                      =                                        γ            ⁡                          (              t              )                                ⁢                                    RTWP              Antenna                        ⁡                          (              t              )                                                            γ            ⁡                          (              t              )                                ⁢                                    N              Antenna                        ⁡                          (              t              )                                          =                                    RoT            Antenna                    ⁡                      (            t            )                          .            
The RoT can hence be measured in the receiver. The major difficulty of any RoT estimation algorithm still remains though, namely to separate the thermal noise power from the interference from neighbor cells. That this is troublesome can be seen from the following equation, where E[ ] denotes statistical expectation, and where Δ denotes the variation around the mean.IN(t)+N(t)=E[IN(t)]+E[N(t)]+ΔIN(t)+ΔN(t).
The fundamental problem can now be clearly seen. Since there are no measurements available in the RBS that are related to the neighbor cell interference, a linear filtering operation can at best estimate the sum E[N(t)]+E[N(t)]. This estimate cannot be used to deduce the value of E[N(t)]. The situation is the same as when the sums of two numbers are available. Then there is no way to figure out the values of the individual numbers. This issue is analyzed rigorously for the RoT estimation problem in T. Wigren, “Soft uplink load estimation in WCDMA”, IEEE Trans Veh. Tech., February, 2009 where it is proved that the noise power floor is not mathematically observable. Nonlinear algorithms that provide approximate estimates of the noise floor are therefore used.
One algorithm that is currently in use estimates the RoT. One main problem solved by the estimation algorithm is the accurate estimation of the thermal noise floor N. Since it is not possible to obtain exact estimates of this quantity due to the neighbor cell interference, the estimator therefore applies an approximation, by consideration of the soft minimum as computed over a relative long window in time. It is important to understand that this estimation relies on the fact that the noise floor is constant over very long periods of time (disregarding the small temperature drift).
The sliding window algorithm described above has the disadvantage of requiring a large amount of storage memory. This becomes particularly troublesome in case a large number of instances of the algorithm are needed, as may be the case when IC is introduced in the uplink. To reduce the memory consumption a recursive algorithm was disclosed in the patent application T. Wigren, “Method and arrangement for memory-efficient estimation of noise floor”, International Patent Application, PCT/SE20061050347, 2006. (P22298). That algorithm reduces the memory requirements of the sliding window scheme discussed above at least by a factor of 100.
The difference with the interference suppressing G-rake receiver as compared to a conventional RAKEreceiver is that each user sees a reduced level of interference, immediately after the so called weight combining step. In G-rake+, a covariance matrix {circumflex over (R)}u, u=1, . . . , U, with the order equal to the number of fingers is first estimated to capture the interference. The codes not used by the present user u may be used in order to estimate {circumflex over (R)}u.
The GRAKE+ receiver uses the estimated covariance matrix that models the interference for computation of the combining weights for the users u, u=1, . . . , U.{circumflex over (R)}uŵu=ĥu, u=1, . . . ,U where ĥu, u=1, . . . , U, is the net channel response of user u and where ŵu are the combining weights.
The effect of the above equation is that GRAKE+ essentially whitens the correlated interference and removes large spectral peaks from interferers at certain finger locations and for certain antenna elements.
Note that GRAKE+ is still a linear receiver. There is a related type of IC receiver for WCDMA which is also linear, denoted the chip equalizer. The difference between GRAKE+ and the chip equalizer is simply the order of certain basic operations.
The now public patent application T. Wigren, “Load estimation in interference whitening systems”, PCT/SE2009/051003 discloses means for estimation of the RoT, as seen by a user after G-rake+. This patent application defines a new signal after G-rake processing and evaluates RoT for that signal.
However, the algorithm of T. Wigren, “Load estimation in interference whitening systems”, PCT/SE2009/051003 requires inversion of the impairment matrix of each user and is too computationally demanding to be preferred presently. The Frequency Domain Pre Equalizing (FDPE) receiver is another interference suppressing receiver. It also affects the measurement of uplink load. The main advantages associated with the structure of this receiver are that the FDPE structure gives significant IS gains and that the FDPE structure achieves IS for all users simultaneously, thereby reducing the computational complexity as compared to the G-rake structure that performs processing individually for all users. Processing blocks are inserted in the uplink receiver structure that is already in place, thereby reducing development cost. The fast Fourier transform (FFT) accelerator hardware developed for LTE can be reused, thereby creating further synergies for the new DUS HW of the RBS.
The FDPE algorithm performs interference whitening in the frequency domain. To explain this in detail, the following time domain signal model can be used
      v    ⁡          (      t      )        =                    ∑                  l          =          0                          L          -          1                    ⁢                          ⁢                        h          ⁡                      (            l            )                          ⁢                  z          ⁡                      (                          t              -              l                        )                                +                            η          v                ⁡                  (          t          )                    .      
Here ν is the received (vector due to multiple antennas) signal, with chip sampling rate, h is the radio channel net response, z is the desired (transmitted) signal, and ηv denotes thermal noise and interference, t denotes discrete time.
Taking the Fourier transform, translates the time domain equation intoV(m)=H(m)Z(m)+N(m)where the quantities are the discrete Fourier transform of the corresponding time domain quantities.
Now a whitening filter can be applied in the frequency domain. It is known that the filter that minimizes the mean square error (the MMSE solution) is given by
            W      MMSE        ⁡          (      m      )        =                              (                                                    R                ^                            d                        ⁡                          (              m              )                                )                          -          1                    ⁢                        H          ^                ⁡                  (          m          )                      =                            (                      [                                                                                                      R                                              0                        ,                        0                                                              ⁡                                          (                      m                      )                                                                                                                                  R                                              0                        ,                        1                                                              ⁡                                          (                      m                      )                                                                                        …                                                                                            R                                              0                        ,                                                                              N                            r                                                    -                          1                                                                                      ⁡                                          (                      m                      )                                                                                                                                                              R                                              1                        ,                        0                                                              ⁡                                          (                      m                      )                                                                                                                                  R                                              1                        ,                        1                                                              ⁡                                          (                      m                      )                                                                                                                                                                                                                                                                                                    ⋮                                                                                                                                          ⋱                                                                                                                                                                                                                R                                                                                                    N                            r                                                    -                          1                                                ,                        0                                                              ⁡                                          (                      m                      )                                                                                                                                                                                                                                                                                                                  R                                                                                                    N                            r                                                    -                          1                                                ,                                                                              N                            r                                                    -                          1                                                                                      ⁡                                          (                      m                      )                                                                                            ]                    )                          -          1                    ⁡              [                                                                                                  H                    ^                                    0                                ⁡                                  (                  m                  )                                                                                                                                              H                    ^                                    1                                ⁡                                  (                  m                  )                                                                                                                                                                                                                                      H                    ^                                                                              N                      r                                        -                    1                                                  ⁡                                  (                  m                  )                                                                    ]            where {circumflex over (R)}d(m) is an estimate of the covariance matrix of V(m). Using a Cholesky decomposition the covariance matrix between the antenna elements can be factored asL(m)·LH(m)={circumflex over (R)}d(m)
The idea behind FDPE is to exploit this factorization and writeWMMSE(m)=(LH(m))−1((L(m))−1Ĥ(m))=Wpre(m)((L(m))−1H(m))so that the desired signal in the frequency domain becomes MMSE pre-equalized in the frequency domain, i.e. given byZpre(m)=Wpre(m)V(m).
This is a user independent processing, which is the same for all users. Hence the wideband received signal is transformed to the frequency domain and the covariance matrix is computed and Cholesky factored, after which the whitened signal is computed. The signal is then transformed back to the time domain where it is further processed for each user. Note that the channels experienced by the RAKE receivers in this processing are obtained from the second factor.
The FDE, Frequency Domain Equalization, algorithm performs equalization and interference suppression in the frequency domain. Contrary to the FDPE, the FDE processing is performed individually for each user. To explain the FDE in, the following time domain signal model is used again
      v    ⁡          (      t      )        =                    ∑                  l          =          0                          L          -          1                    ⁢                          ⁢                        h          ⁡                      (            l            )                          ⁢                  z          ⁡                      (                          t              -              l                        )                                +          i      ⁡              (        t        )              +                            n          thermal                ⁡                  (          t          )                    .      
Here ν is the received (vector due to multiple antennas) signal, h is the radio channel net response, z is the desired (transmitted) signal, i(t) is the interference and nthermal(t) denotes thermal noise, t denotes discrete time.
Taking the Fourier transform, translates the above equation intoV(m)=H(m)Z(m)+I(m)+Nthermal(m)where the quantities are the discrete Fourier transform of the corresponding time domain quantities. Now MMSE equalization can be performed on V(m), separately for each user (different from the FDPE structure). For this purpose, the channel is estimated using the pilot signal, below this fact is emphasized by using the subscriptu for user u. A first method to compute the MMSE filter for the FDE, using time domain calculations is described in E. Dahlman, S. Parkvall, J. Sköld and P. Beming, “3G Evolution—HSPA and LTE for mobile broadband—section 5.1” 2:nd edition, Academic Press, 2008.
However, rather than computing the filter coefficients in the time domain and then transforming to the frequency domain, the MMSE filter coefficients can be directly computed as in T. Wigren, A. KAngas and H. Egnell, “Load estimation in frequency domain pre-equalization systems”, PCT/SE2010/051054.Wu(m)=HuH(m)(HuH(m)HuH(m)+Iu(m)IuH(m)+(Nthermal(m))HNthermal(m))−1, u=1, . . . ,U where the thermal noise power floor matrix estimate, can be obtained by any of the algorithms for noise floor estimation described above, and where Hu(m) is the sampled channel frequency response vector for user u. The use of frequency domain computation is less computationally complex than the method depicted in FIG. 1, and represents the preferred embodiment for implementation of the FDE.
Finally, the equalized signal is computed by a frequency domain multiplication asZFDE(m)=Wu(m)V(m), u=1, . . . ,U after which the inverse FFT is applied to get the signal zFDE,u(t). After this step processing proceeds as in a conventional WCDMA system. The processing is repeated for all users.
In the 3GPPUTRAN architecture, NBAP (Node B Application Part) is the signaling protocol responsible for the control of the Node B by the RNC. RNSAP (Radio Network Subsystem Application Part) is a 3GPPsignaling protocol responsible for communications between Radio Network Controllers. The NBAP and RNSAP protocols allow for signaling of Received total wideband power (RTWF(t)), the estimated thermal noise floor and the received scheduled enhanced uplink power (RSEPS(t)).
The details of the encoding of these messages appear in the specifications 3GPP TS 25.433, UTRAN Ibu Interface Node B Application Part (NBAP) Signaling and 3GPP TS 25.133, Requirements for support of radio resource management.
The signaling breaks the estimated RoT into two pieces, the estimated noise floor and the total wideband power. Note that 3GPP TS 25.433 and 3GPP TS 25.133 state that it is the quantities at the antenna connector that are to be signaled, signaling of other related quantities in these containers represents a proprietary solution.
U-TDOA (Uplink Time Difference of Arrival) is a real time positioning technology for wireless device networks that uses multilateration based on timing of received uplink signals to locate the wireless device.
OTDOA (Observed Time Difference of Arrival) is another real time positioning technology for wireless device networks that uses multilateration based on timing of received downlink signals to locate the wireless device.
The major conceptual difference between UTDOA and OTDOA is that the OTDOA requires multiple transmit points whilst UTDOA utilizes multiple receive points at different locations (typically BS locations), although the position calculation principle is the same.
FIG. 2 illustrates position calculation using the UTDOA method. As illustrated there are three base stations 21 that measure timing of received signals from a wireless device 22.
Assuming that the measurement are successful for the base stations 21, the following relations between the measured TOAs in the base stations 21, the transmission time from the wireless device 22 and the distances between the wireless device 22 and the base stations 21 follow:
                                                           t                              TOA                ,                1                                      +                          b              clock                                =                                    T              transmit                        +                                                                                                r                    1                                    -                                      r                    Terminal                                                                              /              c                                                          ⋮                                                                t                              TOA                ,                n                                      +                          b              clock                                =                                    T              transmit                        +                                                                                                r                    n                                    -                                      r                    Terminal                                                                              /                              c                .                                                        
Here tTOA,i, i=1, . . . , n denotes the measured time of arrivals (TOAs) in the known measuring locations ri, i=1, . . . , n, Ttransmit denotes the transmission time from the wireless device 22 and c is the speed of light. The boldface quantities are the (vector) locations of the base stations 21 and the wireless device 22. bclock denotes the unknown clock bias of the wireless device 22 with respect to cellular system time. Now, in TDOA positioning, time of arrival differences with respect to the own site are formed according to
                                             t                          TDOA              ,              2                                =                                                    t                                  TOA                  ,                  2                                            -                              t                                  TOA                  ,                  2                                                      =                                          T                transmit                            -                              b                clock                            +                                                                                                            r                      2                                        -                                          r                      Terminal                                                                                        /                c                            -                                                                                                            r                      1                                        -                                          r                      Terminal                                                                                        /                c                                                                        ⋮                                                  t                          TDOA              ,              n                                =                                                    t                                  TOA                  ,                  n                                            -                              t                                  TOA                  ,                  1                                                      =                                          T                transmit                            -                              b                clock                            +                                                                                                            r                      n                                        -                                          r                      Terminal                                                                                        /                c                            -                                                                                                            r                      1                                        -                                          r                      Terminal                                                                                        /                                  c                  .                                                                        
In these n−1 equations, the left hand sides are known (with some additional measurement error), provided that the time of transmission difference between the network and UE time can be measured. This is normally achieved with dedicated hardware so called location measurement units (LMUs) or by other procedures. In case of a synchronized network the difference is known. Further the locations of the base stations 21, ri, i=1, . . . , n, can be surveyed to within a few meters and so they are known as well. What remains unknown is the wireless device 22 location, i.e.rTerminal(xTerminalyTerminalzTerminal)T.
In the more common case a two dimensional positioning is performed the unknown position is insteadrTerminal=(xTerminalyTerminal)T.
It then follows that at least three time of arrival differences are needed in order to find a 3D wireless device position and that at least two time of arrival differences are needed in order to find a 2D wireless device position. This, in turn, means that at least four sites need to be detected for 3D wireless device positioning and at least three sites need to be detected for 2D wireless device positioning. In practice, accuracy can be improved if more measurements are collected and a maximum likelihood solution is introduced. There may also be multiple (false) solutions in cases where only a minimum number of sites are detected. The UTDOA method belongs to the set of high precision methods, the inaccuracy is however significantly larger than that of A-GPS. The main advantage of UTDOA is that it provides high precision positioning also indoors, a situation where the availability of A-GPS is very limited.
To perform UTDOA timing measurements also on user data, to increase the signal to noise ratio, one reference receiver de-codes the wireless device signals, and forwards the sequence to cooperating receivers. This procedure is relatively complex and requires a significant amount of signaling. The cooperating receivers are normally located in dedicated hardware close to the positioning node. The decoded reference sequence is used in order to regenerate the transmitted sequence from the wireless device, to allow correlation against each forwarded received set of data from the involved receivers in different locations (typically RBS locations).
The main problem with all terrestrial time difference of arrival positioning methods is to detect/be detected in a sufficient number of non-collocated locations. In the case of UTDOA the problem consists of detection of the same wireless device transmission in a sufficient number of WCDMA base stations (assuming that UTDOA timing measurements are performed in connection to WCDMA RBSs). This is in general a difficult problem since it requires a sufficiently high signal to noise ratio in a number of locations sometimes far away from the wireless device. It needs to be noted that the theoretical minimum of three neighbor locations is not enough in practice. In many situations the number of neighbors may be twice this figure to obtain a reliable performance.
There are however several problems with technology for UTDOA positioning known in prior art.
In case several carriers are available in the base station, it is not known in the positioning node which carrier, if any, UTDOA reference and receivers are available for. Note that UTDOA radio measurements are normally performed in separate HW, therefore it is not evident for which carriers this is possible. In general, having more carriers results in a more expensive radio system.
There is therefore a need for an improved solution for UTDOA positioning which solution solves or at least mitigates at least one of the above mentioned problems.